1
Tests of a proximity focusing RICH with aerogel
as radiator

• Toru Iijima / Nagoya University
• Iijima_at_hepl.phys.nagoya-u.ac.jp

for Belle Aerogel-RICH RD group

• Introduction for aerogel proximity focusing
• Simulation
• Results from the 1st beam test
• Discussion
• Summary

2
Collaborators

• Belle Aerogel-RICH RD group
• I.Adachi2, I. Bizjak5, A. Gorisek5, T. Iijima3,
  M. Iwamoto1, S. Korpar5,8,
• P. Krizan7,5, R. Pestotnik5, M. Staric5, A.
  Stanovnik5,6, T. Sumiyoshi4,
• K. Suzuki2 and T. Tabata1
• 1 Chiba University, Chiba, Japan
• 2 High Energy Accelerator Research Organization
  (KEK), Japan
• 3Physics Department, Nagoya University, Nagoya,
  Japan
• 4Physics Department, Tokyo Metropolitan
  University, Tokyo, Japan
• 5 Jozef Stefan Institute, Ljubljana, Slovenia
• 6 Faculty of Electrical Engineering, University
  of Ljubljana, Slovenia
• 7 Faculty of Mathematics and Physics, University
  of Ljubljana, Slovenia
• 8 Faculty of Chemistry and Chemical Engineering,
  University of Maribor, Slovenia

3
Introduction

• Silica aerogels
• Collidal form of SiO2
• r0.1g/cm3
• Prosity 95 (n1.03)
• n1.0061.06
• ?useful for p/K separation in GeV region
• Applied in many particle/nuclear physics exps.
• But, mostly used in threshold mode (before).
• Limitation for imaging
• Rayleigh scattering
• Cracks

Established technique Simple reliable
Limited momentum coverage Kaon ID in veto mode
4
KEK Aerogel

• Development in 90s for Belle-ACC at the KEK B
  factory.
• New single-step production method developed in
  collaboration
• with Matsushita E.W.Co
• Good optical quality
• Transmission
• Index uniformity
• Long term stability
• Hydrophobicity
• Radiation hardness

Improved optical quality of aerogels ?Break
through for imaging (Aerogel-RICH).
5
Transmission of aerogels

• Almost 3 times better transmission
• (dominated by Rayleigh scattering)

6
Index uniformity

• Dn/(n-1) 3 (FWHM) in Belle-ACC experience

7
Ring viewed by IIT
Yes, rings are there !
8
RICH w/ aerogel radiator

• Larger Dqc between two particle species
• Wider PID coverage ? threshold
• p/K separation
• _at_ higher momentum (10GeV/c?)
• m/p separation
• _at_ low momentum (lt1GeV/c)
• Note p/K sep.(5GeV/c) m/p sep.(1GeV/c)
• Compact ring size
• q 13for n 1.03 / b 1
• Small material thickness
• Photodetection in the visible region
• Short wave length component strongly suppressed
  by Rayleigh scattering.

K-p
p-K
m-p
9
Proximity Focusing ?

• Mirror focusing
• HERMES/LHC-B (open geometry exp.)
• Thick aerogel
• ? Easier to obtain enough light yield.
• Proximity focusing
• Suitable for collider exp.
• Aerogel must be thin enough not to
• deteriorate the angle resolution.
• ? Light yield ?
• Never been considered well by other exps.
• Talk by M.Buenerd (AMS)

Maybe difficult but CHALLENGING ?Our RD
10
Possibility at Future Belle

• Present Belle PID system
• Combination of dE/dx ToF ACC
• KID eff.88/fake8.5.
• Particle ID Holes
• EACC works only for tagging
• e/m-p separation at low momentum
• Proximity focusing aerogel-RICH for the endcap
• to cover give p/K sep.in 0.8ltp(GeV/c)lt4.5
• ?photodetection is hard for Barrel
• photoelectrons must drift perpendicular
• to the B filed axis.
• TOP counter for Barrel ? Talk by T.Ohshima

11
RD History

• By 2000 Some preliminary studies
• Imaging test with IIT / Photodetection /
  Simulation
• busy years for Belle-ACC though,
• March 2001 Proximity focusing aerogel-RICH
  introduced as a possible Belle PID upgrade option
  by T.I.
• J.Stefan institute joined the RD
• Discussion for Super-KEKB/Belle
• June-Oct. 2001 Cosmic ray test at IJS.
• Nov.-Dec. 2001 Beam test at KEK 12 GeV-PS

12
Simulation Ingredients

• Production of Cherenkov photons F-T formula
• Chromatic dispersion Scaled from fused silica
• Rayleigh scattering
• Mean free path based on transmission data
• (with correction of finite acceptance of
  photospectrometer)
• Photodetection
• Bialkari photocathode (QE 25 at 400nm)
• Pad size 5mm/10mm

13
Simulation Result (I)

• Simulated Npe for unscattered photons.
• Normal incidence
• Assume 100 geometrical acceptance for
  photodetection
• Higher n is preferred to have enough light yield
  for pions at 0.8 GeV/c.
• ?Lower n gives better separation at high
  momentum (next slide).

14
Simulation Result (II)

• Simulation indicates
• Npegt12 possible for light velocity particles
• N1.030 gives better separation, but light yield
  at around threshold (p0.8GeV/c) may be critical.
  ?Optimal n 1.05
• Separation _at_ 4GeV/c gt 5s possible even with 10mm
  read-out pad.
• ?Need verification with experiment.

15
1st Beam Test

• Nov.25Dec.2 _at_ KEK-PS, p2 beam line

Aerogel sample
Index Thickness(mm) Lt(mm)
1.029 21 38
1.050 19.5 15
1.070 20.5 7.2
16
Photodetection
Coarse granularity

• 66 Multi-anode PMT array based on
• Hamamatsu R5900-M16
• 36 photocathod coverage
• ?18mm / 30mm pitch
• Two read-out configuration
• Coarse granularity
• Every 4 pads grouped together (144ch)
• ?optimum for light yield measurement
• Fine granularity
• only 1/3 sector connected to TDC (192ch)
• ?optimum for angular reso. measurement

Fine granularity
17
Observed Ring Image

• Coarse granularity, 3GeV/c p-
• n1.029 (2cm thick)

Ring expected from tracking
Hit due to beam
2D hit distribution w.r.t beam
Typical hits / event
18
Data _at_ p 0.5GeV/c

• e-, m-, p- at low momentum with beam line
  polarity at negative

electron veto by gas C
19
Typical distribution

• Ndet sq are obtained by fitting the q
  distribution.
• Gaussian background ? Ns
• Ndet Ns/event
• Also checked Ndet by counting hits in 3s window
  around the peak.
• sq is the single photon angular resolution.

Counting in 3s window
Gaussianbackground
Poisson based on 0 hit bin
Ndet Ns/event
20
Detected photon / event

• Ndet 2.6/2.7 for n1.029/1.050 samples (2cm
  thick).
• Note n1.050 has worse transmission
• Correction by PMT array acceptance applied
  (event-by-event).

2.8cm Novosibirsk
22cm coarse
2cm fine
2cm coarse
21
Angular resolution

• Fine granularity (4.5mm pad)
• ?sq 10mrad at d 20cm, 2cm thick,
  n1.029/1.050.
• d distance between radiator and photodetector
• varied depending on index
• d29cm (1.029), 21cm(1.050), 17cm(1.070)
• Right figure shows the results when normalized to
  d 20cm.

Normalized at d20cm
2cm fine gran.
22
Momentum dependence

• Momentum dependence is as expected.
• Lines in figures are expected dependence fit to
  the data points.
• Effect of multiple scattering at low momentum (sq)

sq (not normalized by d)
Ndet/ring
23
Incident angle dependence

• Almost no dependence on incident angle.

Ndet/ring
sq (not normalized by d)
24
Discussion

• Comparison to expectation (2cm, 3GeV/c pions)

index Ndet (coarse gran.) Ndet (coarse gran.) sq (mrad, fine) sq (mrad, fine)
index Measured Expected Measured Expected
1.029 2.6 2.7 7.0 6.7
1.050 2.7 2.9 9.8 9.7
Assume 36 effective area 70 correction eff.

• Importatnt RD issues
• Improvement of aerogel quality
• Better transmission
• Improvement of photodetection
• Good effective area (36 ? 70)
• Good efficiency for single photon

25
Improvement of aerogels

• Significant increase in Ndet with n1.050
  Novosibirsk sample.
• Ndet 4.6 (28mmt) 2.7 (20mmt) w/ KEK
• 70 increase (40 increase when thickness
  difference accounted).
• We aim at improving the quality at n1.05
• by revisiting
• Production method/procedure
• Precursor material
• Solvent
• Other RD issues
• Size
• Flatness etc.

?Talk by A.F.Danilyuk
26
Photodetection

• Flat Panel PMT array for RD
• Newly developed 8 8 multi-anode PMT by HPK.
• Effective area ?49mm for ?51.7mm package (90
  coverage)
• Single p.e. peak is observable.
• Further performance studies and better
• understanding of the detector behavior.
• Optimization of design
• In real application in 1.5T magnetic field at
  Belle,
• ?Proximity focusing HPD with good effective area

Flat Panel HPD/HAPD?
27
Summary

• Improved optical quality of aerogels in 90s
  opened the window for aerogel-RICH.
• We have tested possibility of proximity focusing
  aerogel-RICH with beam
• Angular resolution lt 10mrad with 5mm pad.
• Light yield 2.7 with photodetection efficiency
  of 25 (3670).
• ? Good starting point
• Results are reasonable and understood.
• They clarify the RD targets.
• Further improvement of aerogel quality (n 1.05)
• Proximity focusing H(A)PD with good effective
  area.

(1.52.0)
gt2
Our hope for improvement factor. If achieved, K/p
separation gt 6s (_at_ 4GeV/c) possible
28
Thank you !!!
29
Backup Slides
30
Lens System

• Optionally used in the beam test.

31
Physics Requirement

• Importance of K/p separation remains unchanged
  also at the Super-Belle.
• Flavor tagging (plt2 GeV/c)
• For any CPV measurement in neutral B.
• Two-body decays (1.5ltplt4GeV/c)
• B?pp/Kp, B?rp(ppp) / Kpp
• B?DK/Dp
• Others hp/hK etc.
• Increased demand at the Super KEKB/Belle
• b?dg/b?sg (required reduction 50?)
• Good separation in inclusive measurements
  (multiple tracks)
• Full reconstruction tag (efficiency/purity)
• How about t / charm ?
• low momentum e/m-p separation (lt 1 GeV/c)
• B?K()ll (b ?sll) (especially in m channel)
• Good K/p lt 5 GeV/c ? good m/p lt 1 GeV/c

Large impact of improved PID
32
Present Belle-PID

• Combination of dE/dx ToF ACC
• Performance
• eff.88/fake8.5.
• Concerns
• Background immunity
• TOF dead time O(10)
• Material thickness
• Radiation hardness
• Particle ID Holes
• EACC works only for tagging
• e/m-p separation at low momentum
• ?Points of improvement _at_ upgrade

33
K/p-ID Impact in b?dg/b?sg

• B ? r g / K g (pp-g /Kp-g)
• B0 ? ppp-p0 g / Kpp-p0 g
• Conditions
• Br(sg) 20Br(dg)
• Present eff/fake from data
• (open histograms)
• Improved eff/fake 0.975/0.025
• 4s separation of two Gaussian
• (hatched histograms)

Present 0.95/0.05 0.975/0.025 0.99/0.01
dg/sg (2chg) 0.64 1.3 (1.1) 2.3 (1.8) 4.6 (2.6)
dg/sg (3chg) 0.45 0.90 (0.78) 1.7 (1.2) 4.2 (1.9)
PID w/ good eff/fake and hermeticity is important
34
Lepton/p-ID Impact in b?sll

• Forward-backward asymmetry (AFB)
• Need to measure AFB as a function of Mll
• May flip sign below and above 1.5GeV
• However,
• In Kmm channel, Mlllt1.5GeV is difficult
• because (almost) no m/p separation.
• Present MUID
• 97.5 _at_ plt1GeV/c by Cherenkov
• ? 2 events in Mll lt 1 GeV
• ? 1.5 events in 1 lt Mll lt 2 GeV

A good Cherenkov detector will give significantly
earlier finding of AFB and test of the sign flip.
35
Flat panel HPD Concept
Photon counting w/ hybrid PD (typical spectrum)
36
Dual Radiator Scheme

• Interesting option, if aerogel transmission is
  improved.
• Light yield can be increased without
  deteriorating the angular resolution.
• Need study effectiveness with simulation